Antenna Noise Temperature Calculator

Rastislav GALUK-OM6AA
1
,Petra GALUKOV, Milo MAZNEK
1
, Pavel
HAZDRA
1
,

Martin MAC
1


1
Dept. of Electromagnetic Field, Czech Technical University, Technick 2, 166 27 Praha 6, Czech
Republic e-mail: om6aa@yahoo.com
2
Dept. of Mathematics and Physics, Charles University in Prague, Ke Karlovu 3, 121 16 Praha 2, Czech
Republic

Abstract: Designers of antennas and their associated components presently enjoy the convenience of
performing fast, accurate, automated, design-related calculations due to the availability of a multitude of
software design tools for modeling electromagnetic fields. Computations involved with specific
problems such as antenna noise temperature, however, require supplementary software tools to augment
the capability of these general purpose state-of- the-art professional programs. To address this need, we
have developed the Antenna Noise Temperature Calculator software program that utilizes either
modeled or measured radiation patterns to calculate antenna noise performance parameters.

1. Introduction

Design objectives for Cosmic-to-Earth radio communication systems mandate a very low
receiver noise budget to permit reception of extremely weak signals. In this respect,
antenna-temperature interaction plays a critical role in the receiving chains noise
performance. The antenna temperature performance calculation is conceptually
described as multiplying the spatial function of antenna gain by the temperature
distribution of the space surrounding the antenna, integrated over all space.

The antenna noise temperature calculation in spherical coordinates is given by the
formula:





(1)



Where

Ta is the antenna noise temperature in a given direction
Tb is the environmental noise temperature
G is antenna gain

Our Antenna Noise Temperature Calculator program (also referred to as ANTC)
computes antenna noise using antenna radiation patterns input from CST Microwave
Studio Software [1], FEKO Software [2] or from an actual antenna measurement.
Antenna noise temperature is calculated for elevation angles from 0 to 90 degrees in
single degree increments. ANTC also enables the calculation of system noise and
average noise temperatures.


2. Axes and Angles Defined

Orientation of axes and angles used with spherical coordinates are shown in Fig.1. The
antenna elevation angle, Alpha, also shown in Fig. 1, ranges from 0 to 90 degrees, from
horizon to zenith respectively. The angles and axes convention for measured data is
shown in Fig. 2. The incremental angle for radiation pattern data calculated by CST and
FEKO software is 1 degree for both Theta and Phi or 1 degree Theta increments for a
maximum of four angles of measured Phi data. The angles and axes convention [3] for
environmental noise temperature is the same as for computed radiation patterns.



Fig. 1 Definition of spherical coordinates (left) and elevation angle Alpha (right)

Fig. 2 Definition of spherical coordinates for measured data
. Main Computation
NTC software, written in C++ programming language, was developed in parallel with
. Read and (if necessary) convert the scale of the radiation pattern and environmental
iation pattern by the function sin(Theta)




3

A
a corresponding MATLAB program to verify the basic claculations. wxWidgets and
pplot libraries were used for the graphical interface and for plotting curves in our C++
program. The program is based on formula (1) and uses the following sequential steps:

a
noise temperature
b. Multiply the rad
c. Calculate radiated power for antenna radiation pattern normalization
the X-axis, using
ental noise temperature by the radiation pattern
is the antenna noise
ome of the more important attributes of the C++ version of ANTC are described in the
.1. Reading the Radiation Pattern and Environmental Noise Temperature
adiation Pattern
put radiation patterns may be obtained from antenna simulation software or from direct

Fig. 3 Input data selector
the "Measured Values" button is selected, the "Select Measured Files" dialog window
definition shown in Fig. 1 left.
d. Rotate the environmental noise temperature by angle Alpha around
precalculated rotation matrices
e. Multiply the rotated environm
f. Integrate results from the previous step over angles Phi and Theta
g. Divide the result of this integration by radiated power. This result
temperature for the appropriate angle alpha

S
following sections.


3

R

In
measurements. The user selects the source of the input radiation pattern (CST, FEKO or
Measured) and the scale type (Linear or Logarithmic) using the input selector switch
shown in Fig. 3. Input files must have the specified data format associated with the
selected file type. This is checked during file reading and an error message appears if an
incorrect format is detected. Since data files from CST MW Studio and FEKO software
contain more data than ANTC needs, the program loads only the necessary data.




If
is opened. See Fig. 4. Here the user loads and may, optionally, modify the measured
data, which are displayed in four tables. At least one and a maximum of four tables must
be populated. For measured data, ANTC calculates a mean value from the number of
loaded radiation patterns. An axially symetric radiation pattern around the Z axis is
created in this manner. The software then tranforms the coordinates according to the


nvironmental Noise Temperature
nds on frequency. Tables of environmental noise
mperature for some important frequencies are included in ANTC. These temperature

Fig. 5
Fig. 4 The measured data dialog window


E

Environmental noise temperature depe
te
tables were created based on ITU documents CCIR rep. 720-2 [4]. The user may select
a frequency-implicit noise temperature table in the noise temperature dialog box (see Fig.
5) or define one's own environmental noise temperatures and input them. The data
format of an input noise temperature file must be the same as the format of tables of
implicit noise temperatures. Environmental noise temperatures for frequencies below 1
GHz are very much affected by cosmic and man-made noise. Consequently, implicit
values for 0.15 GHz and 0.5 GHz are provided as convenient approximations. For better
results, actual measured or defined values should be used.

Environmental noise temperature dialog box



3.2. Enviromental Noise Temperature Matrix Rotation
le alpha, we must rotate the
nviromental noise temperature matrix or antenna radiation pattern through the angle
os(Phi) * sin(Theta)
y = r * sin(Phi) * sin(Theta) (2)
y the Euler rotational matrix around the X axis
ith parameter Alpha. For our definition of angle Alpha (See Fig. 1 right), the Euler

(3)

ation of noise temperature matrix rotation is shown in Fig. 6.

eg.-bottom left, 90-bottom right


To calculate antenna noise temperature for a desired ang
e
Alpha. In ANTC, we choose to rotate noise temperature due to its lower range of
values, in what is from a mathematical point of view an equivalent operation. The first
versions of the C++ program (to version 3.0) employed a direct rotation of the noise
temperature matrix. Since rotation in spherical coordinates is very complicated, the data
format was converted to Cartesian coordinates prior to performing a Cartesian rotation
using the following formulae:

x = r * c

z = r * cos(Theta)

We then multiplied the converted matrix b
w
matrix for rotation around the X axis is:







A 3-D graphical illustr

Fig. 6 Rotation of noise matrix around axis X. 0 deg.-top left, 30 deg.-top right, 60
d


Next, the rotated noise temperature matrix was converted back to spherical coordinates.
he formulae used for converting to spherical coordinates are:
r = sqrt(x+y+z)
mula (2) and reverse
ording to formula (4) are based on circular
nctions, the resulting rotated noise matrix in spherical coordinates has non-linear scales
r angles Theta and Phi. For subsequent calculations it was necessary to find the nearest
where
is the great circle radius
1, 2 are angles the of radiation pattern and noise temperature expressed in
herical coordinates
ting is very time consuming. To increase computation speed,
= | 1 2 | (6)
ce with formula 6, the following, similar,
angles is applied:
)
k, we revised our method of noise matrix rotation. Our
oals were to increase the program speed and enhance performance, reliability and
the original ANTC software was
ing the bi-directional spherical-Cartesian coordinate conversions. Also,
n 0 and 90 degrees. Each element of the basic matrix (not the rotated unit
here) has its own associated index. After applying equations (2), (3) and (4), we find
the appropriate element with appropriate index in the basic matrix for each rotated
T


Phi = atan2(x, y) (4)
Theta = acos( z / r )


Since transformation of the noise matrix according to for
transformation of the rotated noise matrix acc
fu
fo
value of noise temperature for the appropriate angles Theta and Phi of the radiation
pattern. The most accurate calculation is based on finding the great circle minimum
deviation by the formula:


(5)



is the minimum distance deviation
r
1, 2,
sp

Using equation (5) for sor
ANTC sorts based on the following formula:



After sorting data for angles in accordan
formula for

= | 1 2 | (7

Taking into account user feedbac
g
precision. The most critical modification to
streamlin
improvements were made in sorting the rotated matrix values, which previously slowed
calculations.

Our new matrix rotation method is based on angular transformation of a unit sphere in
spherical coordinates. Equations (2), (3) and (4) are applied to a unit sphere for alpha
angles betwee
sp
element of the unit sphere. This is accomplished using formulae similar to (6) and (7).
Thus we create 90 matrix files with a pair of indices in txt format using MATLAB
software. During the calculation, ANTC software selects the corresponding rotation
matrix and assorts actual noise matrix elements according to rotation matrix indices. In
this manner, the noise matrix rotation is transformed into processes of sorting elements.




3.3. Integration of results from previous step over angles Phi and Theta

"Trapezoidal integration," a numerical integration method, is employed to integrate
sults from the previous step over angles Phi and Theta using the following formula:

re


(7)


Since double integration over all space must be performed, all values of items with the
same angle Phi are added together (the first and the last item with the same Phi angle are
added after multiplying these values by ) thereby performing gration over angle
Theta. Then an integration is performed over angle P the last item are
dded after these values are multiplied by ).
hen antenna noise temperature is calculated, a plot with calculated values appears in a
ted noise temperature is displayed as a function of an elevation
ngle. The user may also display the plot by choosing "Tools->Draw Graph of Noise
The displayed plot may be saved as a ".bmp file" by pressing "Save Plot"
the displayed dialog window.
e temperature Ts and system G/Ts ratio are calculated
om antenna noise temperature for a given elevation angle and receiver noise figure. See
inte
hi (the first and
a


4. Additional Features

4.1. Plotting

W
pop-up window. Calcula
a
Temperature."
in

4.2. System Noise Temperature

When antenna noise temperature is calculated, system noise temperature may be
calculated as well. System nois
fr
Fig. 7.

Fig. 7 System noise temperature dialog


4.3. Average Noise Temperature Calculation

The average noise temperature for a given range of elevation angles is a useful parameter
for design and optimization of antennas with variable elevation angle. Prime applications
are antennas intended for receiveing signals from satellites or for EME (Earth-Moon-
Earth communication) use. The elevation angle is determinated by low and high limits
for these applications. Calculated antenna noise temperature is an arithmetic mean value
that corresponds with the angles of this interval.

4.4. Storing Calculated Data

Calculated antenna noise temperature may be saved in text file format. The user selects
"File->SaveOutput" in the main menu.

4.5. Log

ANTC records all operations in a log. The file name of the log is "log.txt", which is
stored in the sotware`s working directory. All operations, whether or not successful, and
calculated data, including intermediate data, are recorded in the log.


5. Practical Application of ANTC


Courtesy of Sergei Zhutyaev RW3BP, we had the opportunity to check the ANTC
software`s accuracy using real-world measured data. Sergei`s EME antenna consists of a
2.4/3.4 m diameter offset dish for the 23 cm band with an extended mesh reflector on half
of the perimeter. The dish was equipped with a dual-mode horn designed by Dmitry
Dmitriev. See Fig. 8. With this antenna system Sergei achieved excellent EME
reception. The best measured G/T ratio based on sun noise vs. cold sky noise
measurement was 19 dB for elevation angles of approximately 30 degrees.

The first step in making the calculated vs. measured data comparison was to model
Sergei`s antenna with CST MW studio software to calculate its radiation pattern
applying the MLFMM method (I-solver). The calculated 3D radiation pattern is shown
in Fig. 9. From this pattern it is apparent that the major portion of the side lobe points
toward the sky, insuring a low antenna noise temperature. The calculated antenna noise
temperature dependence on antenna elevation angle is plotted in Fig.10. Calculated
results for the G/T ratio are shown by the dotted line in Fig. 12.

Since Sergeis intent was to optimize his antenna system for the highest G/T ratio,
Dmitry designed a new feed having a narrower radiation pattern for this purpose. See
Fig.11. The designed system achieved a maximum G/T ratio of about 20dB. Calculated
values for this configuration are shown by the solid line in Fig. 12.







Fig. 8 RW3BP offset dish antenna system



















Fig. 9 Antenna model and its calculated 3D radiation pattern from


Fig. 10 Calculated antenna noise temperature of the antenna shown in Fig.8
CST MW Studio software






Fig. 11 Enhanced Antenna Configuration



Fig. 12 Calculated Antenna G/T Ratio Dependence on Elevation Angle


6. Summary and Conclusion

ANTC software was developed for parabolic dish antenna optimization. Verified with a
real antenna, ANTC used in conjunction with the CST MW Studio program achieves
very good accurancy (similar results are expected, though not yet verified, using FEKO
software). With the current one degree radiation pattern digitization, the program is
able to calculate noise temperature for antennas with gains up to 34 dB. To deal with
very fine digitization and immense data volume requirements associated with higher gain
antennas, the noise analysis should be modified by dividing it into two tasks: an analysis
of the antenna main beam only with very fine digitization followed by an analysis of the
remaining radiation pattern with the current 1 degree digitization. This proposed future
modification of ANTC software would be necessary for calculation of absolute noise
parameters. If a noise relationship study between the feed and parabolic reflector shapes
is required, a scale model should be used for this purpose.

Copies of ANTC software for MATLAB and also the stand-alone C++ version are
available for free download at: http://www. om6aa.eu


Acknowledgements

The authors would like to express their thanks to Robert Valenta for his valuable
technical and language assistance in creating this article. We would also like to thank
Sergei Zhutyaev - RW3BP for providing antenna-measured data and his technical
remarks.



References

[1] Computer Simulation Technology website: www.cst.de
[2] FEKO EM Systems-Software & SS.A. (Pty) Ltd, website: www.feko.info
[3] Masters, Gregory F., Gregson, Stuart F.; Coordinate System Plotting for Antenna
Measurements http://www.nearfield.com/amta/AMTA07-0092-GFM_SFG.pdf
[4] Rep. CCIR, 1990; Rep. 720-2, Fig. 5, 1990.
[5] Ishimaru, Akira; Electromagnetic Wave Propagation Radiation and Scatering Prentice
Hall, New Jersey, 1991. ISBN 0-13-249053-6